ArticlePDF Available

Autism From 2 to 9 Years of Age


Abstract and Figures

Autism represents an unusual pattern of development beginning in the infant and toddler years. To examine the stability of autism spectrum diagnoses made at ages 2 through 9 years and identify features that predicted later diagnosis. Prospective study of diagnostic classifications from standardized instruments including a parent interview (Autism Diagnostic Interview-Revised [ADI-R]), an observational scale (Pre-Linguistic Autism Diagnostic Observation Schedule/Autism Diagnostic Observation Schedule [ADOS]), and independent clinical diagnoses made at ages 2 and 9 years compared with a clinical research team's criterion standard diagnoses. Three inception cohorts: consecutive referrals for autism assessment to (1) state-funded community autism centers, (2) a private university autism clinic, and (3) case controls with developmental delay from community clinics. At 2 years of age, 192 autism referrals and 22 developmentally delayed case controls; 172 children seen at 9 years of age. Consensus best-estimate diagnoses at 9 years of age. Percentage agreement between best-estimate diagnoses at 2 and 9 years of age was 67, with a weighted kappa of 0.72. Diagnostic change was primarily accounted for by movement from pervasive developmental disorder not otherwise specified to autism. Each measure at age 2 years was strongly prognostic for autism at age 9 years, with odds ratios of 6.6 for parent interview, 6.8 for observation, and 12.8 for clinical judgment. Once verbal IQ (P = .001) was taken into account at age 2 years, the ADI-R repetitive domain (P = .02) and the ADOS social (P = .05) and repetitive domains (P = .005) significantly predicted autism at age 9 years. Diagnostic stability at age 9 years was very high for autism at age 2 years and less strong for pervasive developmental disorder not otherwise specified. Judgment of experienced clinicians, trained on standard instruments, consistently added to information available from parent interview and standardized observation.
Content may be subject to copyright.
Autism From 2 to 9 Years of Age
Catherine Lord, PhD; Susan Risi, PhD; Pamela S. DiLavore, PhD;
Cory Shulman, PhD; Audrey Thurm, PhD; Andrew Pickles, PhD
Context: Autism represents an unusual pattern of de-
velopment beginning in the infant and toddler years.
Objectives: To examine the stability of autism spec-
trum diagnoses made at ages 2 through 9 years and iden-
tify features that predicted later diagnosis.
Design: Prospective study of diagnostic classifications
from standardized instruments including a parent inter-
view (Autism Diagnostic Interview–Revised [ADI-R]), an
observational scale (Pre-Linguistic Autism Diagnostic Ob-
servation Schedule/Autism Diagnostic Observation Sched-
ule [ADOS]), and independent clinical diagnoses made
at ages 2 and 9 years compared with a clinical research
team’s criterion standard diagnoses.
Setting: Three inception cohorts: consecutive referrals
for autism assessment to (1) state-funded community au-
tism centers, (2) a private university autism clinic, and
(3) case controls with developmental delay from com-
munity clinics.
Participants: At 2 years of age, 192 autism referrals and
22 developmentally delayed case controls; 172 children
seen at 9 years of age.
Main Outcome Measures: Consensus best-estimate
diagnoses at 9 years of age.
Results: Percentage agreement between best-estimate di-
agnoses at 2 and 9 years of age was 67, with a weighted
of 0.72. Diagnostic change was primarily accounted for
by movement from pervasive developmental disorder not
otherwise specified to autism. Each measure at age 2 years
was strongly prognostic for autism at age 9 years, with
odds ratios of 6.6 for parent interview, 6.8 for observa-
tion, and 12.8 for clinical judgment. Once verbal IQ
(P=.001) was taken into account at age 2 years, the ADI-R
repetitive domain (P=.02) and the ADOS social (P=.05)
and repetitive domains (P = .005) significantly predicted
autism at age 9 years.
Conclusions: Diagnostic stability at age 9 years was very
high for autism at age 2 years and less strong for perva-
sive developmental disorder not otherwise specified. Judg-
ment of experienced clinicians, trained on standard in-
struments, consistently added to information available
from parent interview and standardized observation.
Arch Gen Psychiatry. 2006;63:694-701
usual pattern of develop-
ment beginning in in-
fancy or the toddler years
and defined by deficits in
3 areas: reciprocal social interaction, com-
munication, and restricted and repetitive
While parents typically re-
port concerns in the first year of life,
children do not receive diagnoses until
much later. Several studies have sug-
gested that diagnoses of autism made at
age 2 years are stable through age 3 years,
and diagnoses made by age 5 years are
stable up to late adolescence.
A recent
study reported relatively good diagnostic
stability but limited continuity in symp-
tom severity to age 7 years for children
given autism diagnoses at age 2 years.
Several intervention projects reported di-
agnostic changes and extraordinary levels
of improvement in a substantial minority of
young children with autism.
Other re-
ports found little diagnostic changeand fewer
marked improvements.
Possible expla-
nations for these conflicting results are di-
agnostic instability or the lack of age-
appropriate diagnostic criteria for very young
children. In addition, epidemiological,
and diagnostic studies
have ex-
tended the conceptualization of autism to
include a broader spectrum of disorders that
range from autism to potentially milder forms
of social deficits, including pervasive devel-
opmental disorder not otherwise specified
atypical autism, and As-
perger syndrome.
Recently, investigators
have begun to ask about the stability for the
broader autism spectrum disorder (ASD) as
well as for more narrowly defined autism.
High stability has been found for clini-
cal diagnoses between ages 2 and 3 years
when health care professionals interpreted
standard criteria for autism.
based on the Autism Diagnostic Interview–
Revised (ADI-R), yielding an algorithm op-
erationalizing DSM-IV and International Sta-
tistical Classification of Diseases, 10th
Revision, were not as stable.
At age 2 years,
children with severe retardation were over-
Author Affiliations: University
of Michigan, Ann Arbor
(Drs Lord and Risi); University
of North Carolina, Chapel Hill
(Dr DiLavore); Hebrew
University, Jerusalem, Israel
(Dr Shulman); National
Institute of Mental Health,
Bethesda, Md (Dr Thurm);
University of Manchester,
Manchester, England
(Dr Pickles).
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
diagnosed with autism and children who did not yet show
repetitive behaviors or stereotyped speech were underdi-
Charman and colleagues
found that diagnostic
thresholds from the ADI-R were crossed and recrossed be-
tween ages 2 to 7 years. Moore and Goodson,
using the
ADI-R modified to take into account clinical observa-
tions, found that 88% of children diagnosed with autism
at age 2 years retained that diagnosis at ages 3 and 4 years.
Increases during this period in repetitive behaviors and in-
terests were also found. Stone and colleagues
reported lower
stability for children who initially received diagnoses of
PDD-NOS than autism, though more than 90% of chil-
dren remained within the autism spectrum 1 year later.
The present article reports prospective data from a rela-
tively large sample of autism referrals and a comparison
group of children with developmental delay seen at ages
2, 4 to 5, and 9 years, assessed using standardized in-
struments, including the ADI-R, a structured observa-
tion, and independent clinical diagnoses. Analyses first
addressed the question of diagnostic stability of autism
and PDD-NOS. Because the application of diagnostic mea-
sures to children younger than 3 years is not well estab-
lished, we address the diagnostic utility of the instru-
ments along with changes in the diagnoses of individual
children. A second aim was to identify features at age 2
years that best predicted later diagnosis.
One hundred ninety-two children were prospectively studied
from the time they were referred for evaluation for possible au-
tism before 36 months of age: 111 from North Carolina and 81
from Chicago, Ill. Sample children were consecutive referrals,
seen before 38 months of age, to 4 regional state-funded autism
centers in North Carolina and to a private university hospital
in Chicago. Exclusion criteria included moderate to severe sen-
sory impairments, cerebral palsy, or poorly controlled seizures.
In addition, 22 children with developmental delays between ages
13 and 35 months who met the same exclusion criteria and who
had never been referred for or diagnosed with autism were re-
cruited from the sources of referral to the North Carolina au-
tism centers. Mean (SD) chronological ages at the time of first
assessment for the referred for evaluation groups (North Caro-
lina, 29.2 [4.6] months; Chicago, 29.2 [5.4] months) and the
developmental delay group (26.6 [6.7] months) were not sig-
nificantly different (P=.09). A parent or guardian provided in-
formed consent in accordance with institutional review boards
of the University of North Carolina, Chapel Hill, and the Uni-
versity of Chicago. Assessments were free of charge; feedback
and a report were provided after each assessment.
At approximately age 5 years, 103 North Carolina and 11
Chicago children referred for evaluation and 22 children with
developmental delay were reassessed. At age 9 years, 87 North
Carolina and 68 Chicago children referred for evaluation and
17 children with developmental delay were reassessed, repre-
senting an 80.4% follow-up rate. Attrition was unrelated to origi-
nal diagnosis, sex, verbal or nonverbal IQ, adaptive function-
ing, or language level but was significantly higher for nonwhite
ethnicity. The 172 children with data at both ages 2 and 9 years
form the basis of this report (
Table 1).
Children received a 2-part standard assessment at each point in
the study. Most frequently, parents were interviewed at home and
then the child and family were seen for a second session at the
child’s school or clinic. The Vineland Adaptive Behavior Scales,
a standardized measure of adaptive functioning based on a par-
ent interview, were administered immediately following the ADI-R
at each age. At age 2 years , all but 1 child (given the Stanford-
Binet), were administered the Mullen Scales of Early Learning.
At age 9 years, the selection of cognitive tests followed a stan-
dard hierarchy designed for use when children could not achieve
a basal score or achieved ceiling scores: 39 children, Wechsler
Intelligence Scale for Children
; 73 children, Differential Abil-
ity Scales
; 51 children, Mullen Scales of Early Learning; and 6
children, other. Because raw scores frequently fell outside stan-
Table 1. Descriptive Characteristics by Best-Estimate Diagnoses at Ages 2 and 9 Years in 172 Children
Diagnosis, 2 y
Diagnosis, 9 y
(n = 84)
(n = 46)
(n = 42)*
(n = 100)
(n = 35)
(n = 37)*
Female, % 14 11 40 14 14 41
White, %† 68 72 76 65 83 76
African American, % 30 26 21 31 17 24
Age, mo, at baseline assessment at 2 y, mean (SD) 29.1 (4.7) 29.1 (5.6) 28.8 (5.5) 29.0 (4.9) 30.3 (5.3) 27.8 (5.5)
Age, mo, at follow-up at 9 y, mean (SD) 110.1 (15.7) 113.8 (17.1) 114.9 (11.8) 111.5 (16.5) 111.1 (15.8) 115.6 (11.0)
Limited speech, 2 y, %‡ 74 57 50 74 51 46
Limited speech, 9 y, % 30 4 7 28 0 6
2 y, mean (SD) 61.0 (12.3) 64.1 (10.3) 65.7 (9.4) 61.0 (12.3) 64.7 (8.2) 66.7 (9.7)
VABC, 9 y, mean (SD) 43.7 (22.8) 57.4 (26.6) 58.2 (27.4) 39.3 (18.6) 69.4 (22.6) 63.9 (29.1)
Ratio verbal IQ, 2 y, mean (SD) 26.4 (15.3) 45.6 (21.7) 57.9 (23.9) 28.5 (16.7) 49.6 (24.3) 58.5 (22.2)
Ratio verbal IQ, 9 y, mean (SD) 41.2 (36.5) 71.7 (36.9) 60.4 (31.1) 35.1 (26.8) 91.5 (32.2) 69.7 (33.4)
Ratio nonverbal IQ, 2 y, mean (SD) 63.3 (16.9) 74.0 (22.3) 72.7 (26.5) 62.7 (19.2) 80.0 (20.0) 73.1 (23.8)
Ratio nonverbal IQ, 9 y, mean (SD) 54.0 (30.9) 75.4 (33.3) 67.9 (33.5) 50.5 (28.9) 88.9 (25.5) 72.7 (34.3)
Abbreviations: ADI-R, Autism Diagnostic Interview–Revised; ASD, autism spectrum disorder; PPD-NOS, pervasive developmental disorder not otherwise
specified; VABC, Vineland Adaptive Behavior Composite.
*The nonspectrum group includes all of the children with developmental delay as well as children referred for evaluation who did not receive ASD diagnoses.
†Four children were of mixed or Hispanic ethnicity.
‡Defined as a score of 2 on the ADI-R overall level of language (5 words used on a daily basis).
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
dard ranges for deviation scores, ratio IQs were calculated sepa-
rately for verbal and nonverbal subtests.
Three measures of diagnosis were obtained at ages 2 and 5
years. Before the direct assessment, a research associate admin-
istered to parents a toddler version of the ADI-R, which in-
cluded additional questions about early development and symp-
tom onset.
The toddler ADI-R is a standardized semistructured
interview of 132 questions. It yields a diagnostic algorithm for
autism by providing scores in 3 domains, social reciprocity, com-
munication, and restricted, repetitive behaviors, and has items
about age at onset. Adequate validity and interrater and test-
retest reliability have been established for children from age 3
years to adults.
For the purpose of these initial analyses, PDD-
NOS was defined post hoc as not meeting autism criteria on
the ADI-R but falling within 1 to 2 points of autism cutoffs for
algorithm criteria in the social and/or communication do-
mains, with no requirement for repetitive behavior.
diately after conducting the interview, the research associate
dictated a 2-page summary, without scoring the algorithm or
referring to individual scores. This text was used in the con-
sensus diagnosis at ages 2 and 5 years.
The Autism Diagnostic Observation Schedule (ADOS)
and an adaptation for younger children, the Pre-Linguistic Au-
tism Observation Schedule (PL-ADOS),
provided standard-
ized observation of social and communicative behavior. In 1999,
the PL-ADOS and the former ADOS
were combined into a
single instrument with separate modules for children at differ-
ent language levels. The algorithm for the ADOS uses thresh-
olds in social reciprocity and communication domains, as well
as an overall cutoff. Reliability and validity have been estab-
lished for children as young as 2 years.
Cutoffs for autism pro-
vide clear differentiation between children with autism and ver-
bally matched children with nonspectrum disorders. However,
the overlap between the narrower classification of autism and
the broader classification of ASD is considerable.
We refer to
the administered test as the PL-ADOS because it included ad-
ditional tasks and scores not retained in the ADOS module 1,
but the ADOS algorithm was used for analyses.
At initial assessment, a PL-ADOS (n=172) was administered
to all subjects referred for evaluation for autism and with devel-
opmental delay. At age 5 years, the PL-ADOS (n=119) or ADOS
module 2 (n=11) was administered. At age 9 years, the ADOS
modules 1 (n =64), 2 (n =46), and 3 (n= 60) were administered.
The ADI-R and PL-ADOS/ADOS items were scored during ad-
ministration; algorithms were completed after the clinical diag-
nosis was made and did not yet exist when the children were age
2 years. Both the ADI-R and PL-ADOS provide item totals for so-
cial, communication (for the ADI-R, nonverbal communication
was used here), and repetitive-behavior domains.
Clinical diagnoses were made at ages 2, 5, and 9 years, using
somewhat different procedures. For the 2-year-olds, following
psychological assessment, 2 clinicians reviewed all test results
and the ADI-R summary, discussed the content of the PL-ADOS,
and proposed a binary clinical diagnosis (autism, not autism)
to which they applied a certainty rating that generated an au-
tism spectrum score from 1 (certain not autism) to 10 (certain
autism). There was no attempt to train the clinicians, who were
clinical and educational psychologists, in making standard di-
agnoses of 2-year-olds. Certainty scores were initially introduced
because clinicians were uncomfortable making diagnostic de-
cisions for such young children. For purposes of analysis, cer-
tainty scores were grouped into definite nonspectrum (1 and 2),
ASD including PDD-NOS and less certain cases of atypical au-
tism (3-7), and definite autism (8-10). This approach confounds
certainty with severity in that PDD-NOS by definition involves
less comprehensive and/or less intense symptoms. As present-
ed in
Table 2, unsurprisingly, children described as having PDD-
NOS received lower scores on diagnostic measures, indicating
fewer or less severe symptoms.
One examiner carried out the assessment at age 5 years for
each child and followed the procedures described earlier to make
a clinical diagnosis. In about two thirds of cases, examiners were
unfamiliar with the child. For the 9-year-olds, most cases were
seen by 2 examiners, both unfamiliar with the child: 1 for the
ADI-R/Vineland Adaptive Behavior Scales and one for the ADOS
and psychometrics. The clinical diagnosis was made jointly.
For the best-estimate diagnoses at both 2 and 5 years of age,
2 psychologists considered the independent clinical diagno-
sis, the ADI-R and ADOS algorithm scores, and the cognitive,
language, and adaptive test scores. They read the ADI-R notes,
watched the PL-ADOS/ADOS videotape, and discussed all the
findings from that age until they reached a consensus. Follow-
ing DSM-IV, distinctions between autism and PDD-NOS were
made on the basis of number of domains affected as well as the
Table 2. ADI-R and ADOS Scores by Initial Best-Estimate Diagnoses at Ages 2 and 9 Years*
Diagnosis, 2 y
Diagnosis, 9 y
(n = 84)
(n = 46)
(n = 42)
(n = 100)
(n = 35)
(n = 37)
ADI-R social domain, 2 y 19.7 (4.2) 14.7 (5.7) 9.7 (5.8) 18.6 (5.2) 15.0 (5.1) 9.6 (6.3)
ADI-R social domain, 9 y 25.0 (5.5) 20.8 (7.1) 13.5 (9.0) 25.4 (4.1) 18.7 (8.0) 11.5 (8.2)
ADI-R nonverbal communication domain, 2 y 10.0 (2.0) 8.3 (3.0) 5.8 (3.5) 9.8 (2.1) 7.5 (2.9) 5.9 (3.8)
ADI-R nonverbal communication domain, 9 y 11.8 (2.9) 8.8 (3.9) 5.4 (4.0) 11.8 (2.4) 7.4 (4.0) 4.7 (4.0)
ADI-R repetitive domain, 2 y 4.1 (1.5) 3.1 (2.3) 2.2 (1.7) 4.0 (1.8) 3.3 (1.9) 1.7 (1.3)
ADI-R repetitive domain, 9 y 5.9 (2.6) 5.5 (3.2) 3.8 (2.8) 6.3 (2.5) 4.9 (3.1) 4.7 (4.0)
ADOS social domain, 2 y 12.6 (1.7) 8.8 (3.4) 4.6 (3.6) 11.6 (3.1) 8.9 (3.4) 4.9 (3.7)
ADOS social domain, 9 y 10.3 (3.0) 7.1 (3.8) 5.0 (3.6) 10.7 (2.3) 5.6 (3.0) 3.6 (3.1)
ADOS communication domain, 2 y 6.5 (1.4) 4.4 (1.8) 2.5 (2.2) 5.9 (1.8) 5.0 (2.4) 2.5 (2.1)
ADOS communication domain, 9 y 6.4 (2.0) 4.8 (2.4) 3.5 (2.2) 6.8 (1.7) 3.9 (1.5) 2.6 (2.0)
ADOS repetitive domain, 2 y 4.0 (1.5) 2.5 (1.4) 0.8 (1.0) 3.6 (1.7) 2.3 (1.5) 1.0 (1.4)
ADOS repetitive domain, 9 y 2.9 (2.1) 1.7 (1.7) 1.3 (1.3) 3.1 (1.9) 1.0 (1.0) 1.0 (1.0)
Abbreviations: ADI-R, Autism Diagnostic Interview–Revised; ADOS, Autism Diagnostic Observation Schedule; PPD-NOS, pervasive developmental disorder not
otherwise specified.
*Values are expressed as mean (SD). The ADOS scores for age 2 years used the module 1 algorithm. At age 9 years, for comparability across modules, all
ADOS scores were converted to module 2 (see Lord et al
for ranges). The ADI-R totals include “ever” scores. The nonspectrum group included all children with
developmental delay.
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
intensity and number of symptoms; clinical certainty ratings
were taken into account but it was left to the clinicians to de-
cide how to use information about a particular child. Parallel
information for age 9 years was used to generate a consensus
best-estimate diagnosis by an independent psychologist and child
psychiatrist blind to earlier diagnoses.
Reliability was initially obtained on the diagnostic mea-
sures (ADI-R, PL-ADOS, and ADOS) after intensive training
until each pair of examiners reached more than 90% exact agree-
ment (␬⬎0.70) on individual items for the ADI-R and 80%
exact agreement (␬⬎0.60) on codes for the PL-ADOS/ADOS
for 3 consecutive administrations. Approximately every sixth
administration of each instrument was scored by 2 raters, yield-
ing between 0.60 and 0.80. Reliability for clinical diagnoses
at age 2 years was measured in 1 in 6 cases with 92% agree-
ment for autism/not autism. The intraclass correlation for cer-
tainty ratings was 0.89. For clinical diagnoses at ages 5 and 9
years, agreement between the examiners was established on cases
outside this study and monitored once a month (overall agree-
ment 90% for best-estimate autism cases, and 83% for chil-
dren with PDD-NOS and nonspectrum disorders).
Diaries completed by parents summarized educational and other
treatments their children had received during each year. Two rat-
ers coded the diaries, having first established reliability on gen-
eral classifications (eg, 1 to 1 vs group). There was considerable
variation in type and amount of treatment. For the purposes of
this article, treatment intensity was defined very crudely by hours
of treatment (including education and formal homeprogramming).
All analyses were undertaken in Stata 8.0.
Agreement among con-
temporaneous diagnostic measures and between baseline and
follow-up diagnosis was assessed using statistics that correct for
chance agreement for nominal measures.
Prediction of autism
and ASD used logistic regression. To compare odds ratios (ORs)
we used Wald tests of interactions from a 2-response generalized
estimating equations logistic model with an exchangeable work-
ing correlation matrix and robust parameter covariance matrix.
To assess the effect of treatment, there was a need to take
account of children’s differential access to treatment.
To con-
trol for such selective treatment assignment, an instrumental
variable approach was used, requiring identification of a vari-
able that, while associated with treatment received, was as-
sumed, given treatment (and confounders), unrelated to out-
Recruitment site (North Carolina or Chicago) was used
as an instrumental variable approach.
Table 1 and Table 2 describe the sample by initial and fol-
low-up best-estimate diagnoses. Rates of diagnosis of au-
tism (and autism plus PDD-NOS) were 55% (81%) for the
ADI-R, 65% (83%) for the PL-ADOS, 38% (69%) for the
clinicians, and 49% (76%) according to the best-estimate
diagnosis. Percentage agreement () was 85.5% (0.53) for
interview-observation, 81.7% (0.47) for interview-
clinician, and 84.3% (0.53) for observation-clinician.
In contrast to the ADI-R and the PL-ADOS,
Figure 1
shows that clinicians rarely (2 in 172 cases or 1%) classi-
fied children as having autism who had not been classi-
fied in the same way by 1 of the other measures. On the
other hand, clinicians relatively frequently (26 in 172 cases
or 15%) indicated autism as not present when both inter-
view and observation classified it as present, though in 19
(73%) of these cases the clinician indicated PDD-NOS. Not-
withstanding, best-estimate autism prevalence was consis-
tently high among children identified by clinicians.
For ASD diagnoses, Figure 1 and
Table 3 show that
the ADI-R and PL-ADOS had similar levels of inclusion,
with both more inclusive than clinical judgment. Levels of
agreement with the contemporaneous best-estimate diag-
nosis, reflecting the relative weight attached to each mea-
sure in coming to the best-estimate diagnosis at age 2 years,
were 0.40 for the interview, 0.54 for the observation, and
0.67 for the clinical judgment (of 1.00 maximum).
Overall Prevalence, 49%
23 (17%)
26 (58%)
16 (6%)
42 (2%)Clinician
11 (100%)
51 (98%)
Overall Prevalence, 76%
7 (29%)
21 (52%)
9 (0%)
2 (100%)
17 (6%)
2 (50%)
7 (100%)
107 (99%)
1 (100%)
2 (50%)
Figure 1. Frequency of diagnostic combinations and contemporaneous best-estimate diagnosis prevalence (in parentheses) at age 2 years. A, Autism.
B, Autism spectrum. PL-ADOS indicates Pre-Linguistic Autism Diagnostic Observation Schedule; ADI-R, Autism Diagnostic Interview–Revised.
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
Table 3 shows that, according to the best-estimate diag-
nosis, between ages 2 and 9 years the proportion with
autism increased from 49% to 58%, mainly because fewer
children were classified as having PDD-NOS. The best-
estimate diagnosis improved between ages 2 and 9 years
for 18 children (8%) (only 1 from autism to nonspectrum
disorder), compared with 38 (18%) with worse classifica-
tion. Overall exact agreement between the best-estimate di-
agnoses at ages 2 and 9 years was 67% (=0.47), 76% for
autism vs nonautism (=0.51), and 90% for autism spec-
trum vs nonspectrum ( =0.72). For 112 children as-
sessed at age 5 years, stability was 72% ( =0.72) from ages
2 to 5 years and 88% ( =0.92) from ages 5 to 9 years.
Figure 2 and Table 3 also show the relative perfor-
mance of individual and combinations of measures at age
2 years in predicting the best-estimate diagnosis at age 9
Table 3. Cross-tabulation of Initial Diagnostic Measures and Best-Estimate Diagnoses at Ages 2 and 9 Years*
Age 2 Diagnostic
Measure No. (%)†
Best-Estimate Diagnosis, 2 y
Best-Estimate Diagnosis, 9 y
Autism PDD-NOS Nonspectrum Autism PDD-NOS Nonspectrum
Autism 94 (55) 67 18 9 73 16 5
PDD-NOS 45 (26) 15 19 11 20 14 10
Nonspectrum 33 (19) 2 9 22 7 5 21
Autism 111 (65) 80 24 7 82 22 7
PDD-NOS 31 (18) 3 19 9 14 8 9
Nonspectrum 30 (17) 1 3 26 4 5 21
Autism 65 (38) 63 2 0 58 6 1
PDD-NOS 53 (31) 19 32 2 32 16 5
Nonspectrum 54 (31) 2 12 40 10 13 31
Autism 84 (49) 71 12 1
PDD-NOS 46 (27) 27 14 5
Nonspectrum 42 (24) 2 9 31
No. (%)‡ 102 (49) 59 (28) 53 (25) 100 (58) 35 (20) 37 (22)
Abbreviations: See Table 2.
*Values are expressed as number of children unless otherwise specified. The nonspectrum group consists of all children with diagnoses other than autism
spectrum disorder. This includes all of the children initially seen in the developmental delay group, as well as some children referred for evaluation.
†Number and percentages of children seen at age 2 years and at age 9 years.
‡Number and percentages of all children seen at age 2 years and number and percentages of children seen at age 9 years.
Overall Prevalence, 58%
23 (43%)
26 (65%)
16 (56%)
42 (14%)Clinician
11 (82%)
51 (90%)
Overall Prevalence, 78%
7 (43%)
21 (67%)
9 (44%)
2 (100%)
17 (12%)
2 (50%)
7 (86%)
107 (96%)
1 (100%)
2 (100%)
Figure 2. Frequency of diagnostic combinations at age 2 years and prevalence of best-estimate diagnosis (in parentheses) at age 9 years. A, Autism.
B, Autism spectrum. PL-ADOS indicates Pre-Linguistic Autism Diagnostic Observation Schedule; ADI-R, Autism Diagnostic Interview–Revised.
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
years. Classifications of autism were frequent for all cli-
nician-positive combinations of measures. The measure
of clinical diagnostic uncertainty at age 2 years was
strongly associated with change. While just 10% of chil-
dren with definitely nonspectrum diagnoses and 18% of
the children with definite autism changed diagnosis, 43%
of the children with less certain diagnoses changed clas-
sification. Each instrument was strongly prognostic for
autism with an OR of 6.6 (95% confidence interval [CI],
3.3-12.9) and sensitivity of 73% and specificity of 71%
for the ADI-R; OR of 6.8 (95% CI, 3.4-13.5) with sensi-
tivity of 82% and specificity of 60% for the PL-ADOS/
ADOS; and OR of 12.8 (95% CI, 5.3-30.8) with sensitiv-
ity of 58% and specificity of 90% for clinical judgment.
In a simple additive logistic regression for best-
estimate autism diagnosis at age 9 years, all 3 diagnostic
measures at age 2 years made an independent contribu-
tion to prediction, with a partial OR of 3.4 (95% CI, 1.6-
7.3) (P=.001) for the ADI-R; partial OR of 2.4 (95% CI,
1.0-5.3) (P=.04) for the PL-ADOS/ADOS, and partial OR
of 6.2 (95% CI, 2.4-16.2) (P=.001) for clinical diagno-
sis, giving an overall sensitivity of 75% and specificity of
78%. Similar analyses showed the ADI-R domain scores
at age 2 years made independent contributions (social,
P=.07; communication, P= .01; repetitive, P=.03). When
verbal IQ (P.001) and nonverbal IQ (P.60) at age 2
years were covaried (lower verbal IQ increased the odds
of autism), only the ADI-R repetitive domain remained
significant (social, P = .30; communication, P =.40; re-
petitive, P=.02). For the PL-ADOS at age 2 years, inde-
pendent prediction from social and repetitive domains
(social, P =.003; communication, P =.90; repetitive,
P=.002), while reduced, remained significant (social,
P=.05; communication, P=.30; repetitive, P= .005) in the
presence of verbal (P= .01) and nonverbal (P = .90) IQ.
Tests comparing the ORs for predicting autism and
ASDs showed some specific relationships with instru-
ments and domains. While nonverbal IQ at age 2 years
did not predict autism at age 9 years, higher nonverbal
IQ and higher PL-ADOS/ADOS communication scores
predicted ASD diagnoses (interactions, P= .006 and P.03,
respectively). The ADI-R repetitive score at age 2 years
predicted ASD at age 9 years more strongly than it pre-
dicted autism (interaction, P= .006).
As expected by their definition, the mean “most abnor-
mal 4 to 5” or “ever”/lifetime ADI-R algorithm scores in
Table 2 are higher at age 9 years than age 2 years. By con-
trast, the mean ADI-R total score based on current items
(excluding verbal items) indicated a marked reduction (8.1
points [95% CI, 6.4-9.7]; P.001) in abnormality, and PL-
ADOS/ADOS scores (corrected for the number of pos-
sible items in the algorithm and the distribution of social
and communication items) also fell (2.1 points [95% CI,
3.2-1.0]; P.001). Change-score analysis of ADI-R and PL-
ADOS/ADOS item totals gave similar findings, with no sig-
nificant associations with sex (P=.70 and .30), ethnicity
(P=.30 and .50), mother’s education (P= .40 and .30) nor
baseline verbal (P=.10 and .07) or nonverbal (P=.20 and
.50) IQs or adaptive behavior (P=.50 and .70).
This improvement contrasted with a marked worsen-
ing during the same period in mean adaptive-behavior stan-
dard scores from 63 to 51 (−12.1 points [95% CI, 15.9-
8.4]; P.001). The decline was associated with low verbal
(P.001) and nonverbal (P.001) IQ at age 2 years and
high ADI-R symptom severities in the social (P.001) and
nonverbal communication (P.001) domains at age 2 years
but not with restricted and repetitive behavior (P=.30).
Change in adaptive behavior was not associated with eth-
nicity (P=.10), sex (P=.30), or mother’s education (P=.60).
Vineland correlations from ages 2 to 5 years were 0.72; from
age 5 to 9 years, 0.85; and from ages 2 to 9 years, 0.62. This
decline in functioning is also evident from Table 1. While
all 3 groups had similar functioning at age 2 years, the au-
tism group at 9 years of age had markedly lower scores.
Table 1 suggests a quite distinctive profile for the PDD-
NOS group at age 9 years, with markedly higher verbal IQ
and, to a lesser extent, nonverbal IQ compared with dif-
ferences in group means at age 2 years.
For each ADI-R and PL-ADOS domain score, regression
prediction of each domain score at age 9 years by the set
of 3 domain scores at age 2 years showed significant con-
tinuity within the same domain. The 1 exception was the
ADOS communication score at age 9 years that was pre-
dicted by the ADOS social (P=.01) and repetitive (P=.002)
domains at age 2 years, with no significant independent con-
tribution from communication (P=.70). Other indepen-
dent cross-domain predictions occurred for the PL-ADOS
social score at age 2 years, predicting the repetitive do-
main score at age 9 years (P=.008), and for the ADI-R, where
nonverbal communication score at age 2 years indepen-
dently predicted social scores at age 9 years (P=.02) and
social scores at age 2 years independently predicted non-
verbal communication scores at age 9 years (P=.003).
Our rather crude measure of hours of treatment was as-
sociated with worsening of the ADI-R total score (P=.01),
adaptive behavior (P.001), and PL-ADOS/ADOS total
score (P=.06). However, this did not take into account
selective treatment exposure, which was strongly asso-
ciated with region of referral (P=.003). Using region as
an instrument for treatment exposure in a 2-stage least
squares regression did not alter the estimated direction
of effects, but all effects were then nonsignificant (P=.08,
.10, and .08, respectively).
Diagnosis of autism in 2-year-olds was quite stable up
through 9 years of age, with the majority of change asso-
ciated with increasing certainty of classifications moving
from ASD/PDD-NOS to autism. Only 1 of 84 children with
best-estimate diagnoses of autism at age 2 years received a
nonspectrum diagnosis at age 9 years, and more than half
of children initially diagnosed with PDD-NOS later met au-
tism criteria. Nevertheless, more than 10% of children with
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
diagnoses of PDD-NOS at age 2 years received nonspec-
trum best-estimate diagnoses (ie, not autism or ASD) by
age 9 years, and nearly 30% continued to receive diag-
noses of PDD-NOS, indicating mild symptoms at age 9 years.
A significant minority of children with milder difficulties
within ASD at age 2 years showed only mild deficits in the
clinical ASD range at age 9 years. Classifications changed
substantially more often from ages 2 to 5 years than from
ages 5 to 9 years. The bulk of change in diagnosis occur-
ring in early years is consistent with another recent study.
At age 2 years, diagnostic groups were more similar in func-
tioning and IQ than the diagnostic groups identified at age
9 years, when the autistic group showed very poor adap-
tive functioning and the PDD-NOS group, much less ab-
normal verbal and nonverbal IQ.
Among this specialized group of clinicians, clinical
judgment of autism at age 2 years was a better predictor
of later diagnosis than either standardized interview or
observation. Contemporaneous agreement between clini-
cal judgment and best-estimate judgment for 2-year-
olds was equal to that found between experienced raters
in the DSM-IV field trials for older children and adults.
Though the clinical diagnoses at age 2 years were made
without knowledge of the ADI-R and ADOS algorithm
scores, each clinician had administered either the PL-
ADOS or the ADI-R and had the opportunity to discuss his
or her impressions with the experienced clinician who had
administered the other instrument. Thus, the information
available to them was very different from the information
obtained during a typical single office visit to a clinical psy-
chologist or developmental pediatrician. The use of stan-
dardized measures seems likely to have improved the sta-
bility of diagnosis both directly through straightforward use
of algorithms for autism and ASD and also indirectly through
structuring clinical judgment. Of cases in which the clas-
sifications yielded by both instruments were not sup-
ported by the clinicians at age 2 years, 40% were children
with severe mental retardation (and not autism) or chil-
dren with very difficult behavior (and not autism), while
the remainder were mild cases of autism characterized as
uncertain. On the other hand, clinical judgments were con-
sistently underinclusive at age 2 years, both for narrow di-
agnoses of autism and for broader classifications of ASD
at age 9 years. Thus, scores from standardized instru-
ments also made real contributions beyond their influ-
ence on informing and structuring clinical judgment. Over-
all, while standardized research instruments at age 2 years
did not fully capture the insight in the form of certainty
ratings made by experienced, well-trained clinicians, this
insight was not by itself sufficient.
A positive ADI-R or PL-ADOS/ADOS classification of
autism or PDD-NOS, when contradicted by the other mea-
sures, was of limited prognostic value. Nonetheless, both
instruments and clinical judgment added to the predic-
tion at age 9 years. The independent predictive power of
the communication domain in the PL-ADOS/ADOS and
both the social and communication domains in the ADI-R
was modest, standing in contrast with the PL-ADOS/
ADOS social and both ADI-R and PL-ADOS/ADOS repeti-
tive domains, which made independent contributions, simi-
lar to the findings of Charman and colleagues.
These and
other findings support the conceptualization of ADI-R and
ADOS social and nonverbal communication items as re-
flecting 1 factor. The limitations of the repetitive domain
score of the PL-ADOS/ADOS, based on a brief sample of
behavior, are well understood,
and several studies have
found that a significant number of children who receive
autism diagnoses in later preschool years are not de-
scribed as having repetitive behaviors before 30 months of
To find the repetitive domain score from the ADI-R
and the PL-ADOS/ADOS so strongly predictive of progno-
sis for autism and ASD 7 years later, both before and after
verbal IQ was taken into account, was surprising. As ex-
pected, low verbal IQ was also associated with increased
probability of an outcome of autism or ASD.
As a group,
children with uncertain clinical diagnoses and high ver-
bal and nonverbal IQs at age 2 years who showed more
prosocial behavior (a relatively low social score on the
ADOS) and little or no repetitive behavior during the ADOS
and ADI-R were most likely to change diagnosis from au-
tism to PDD-NOS and PDD-NOS to nonspectrum catego-
ries at age 9 years and were least likely to show losses in
adaptive behavior at age 9 years (and so have relatively bet-
ter outcome in everyday skills).
As reported elsewhere,
the overall totals on the ADI-R
and ADOS were not systematically related to change in au-
tistic symptoms from age 2 to 9 years. The lack of evidence
for a true association between the amount of therapeutic
intervention and amount of diagnostic change is not encour-
aging for very time-intensive treatments but may reflect our
rather gross quantitative measure of hours of intervention,
which had no control for kind or quality of treatment.
This study has the usual strengths and limitations of a
prospective cohort study. Children were identified at young
ages, which allowed for prospective study but also meant
that these cohorts are not necessarily representative of chil-
dren referred for autism at older ages. The oldest of these
children was referred 14 years ago, which also means that
a cohort of 2-year-olds today might be rather different. The
clinicians providing the clinical judgments were very ex-
perienced clinicians, though not with 2-year-olds, who made
up a relatively small proportion of routine referrals at that
time. This lack of familiarity with 2-year-olds likely con-
tributed to the clinicians’ consistently underinclusive judg-
ments, a finding replicated by others,
which deserves spe-
cial attention at a time when most concern is about
overdiagnosis of ASD for older children.
Overall, referrals of 2-year-olds for possible autism to
2 very different programs in different regions (North Caro-
lina and Chicago) included many more children who ac-
tually had ASD than we expected, with just less than half
of the referred children receiving autism diagnoses and
75%, ASD diagnoses. This attests to the ability of com-
munity physicians, and the parents who for the most part
initiated the process, to make appropriate referrals when
a free evaluation was easily accessible, though it is im-
portant to remember that we cannot determine how many
children were not referred who should have been.
In turn, clinicians in the study, using standardized in-
struments and their own judgments to integrate informa-
tion into a best-estimate diagnosis at age 2 years, were able
to make classifications that predicted diagnosis within the
autism spectrum at age 9 years for almost all cases. There
are real questions about the usefulness of PDD-NOS as a
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
categorical diagnosis. However, especially for very young
children, having a way for experienced clinicians to ac-
knowledge their uncertainty about some 2-year-olds was
ultimately helpful as a means of flagging children who by
age 9 years had a range of difficulties from autism to very
mild social deficits. On a more somber note, because more
than half of the children with PDD-NOS clinical diag-
noses at age 2 years received best-estimate diagnoses of au-
tism by age 9 years, health care professionals should be wary
of telling parents that their young children do not have au-
tism, only PDD-NOS. In the end, the development of mean-
ingful measures of continuous dimensions of behavior in
ASD should improve research and practice.
Submitted for Publication: June 6, 2005; final revision re-
ceived November 23, 2005; accepted December 21, 2005.
Correspondence: Catherine Lord, PhD, University of
Michigan Autism and Communication Disorders Cen-
ter, 1111 E Catherine St, Ann Arbor, MI 48109 (celord
Financial Disclosure: Drs Lord and Risi receive royal-
ties from the publication of the Autism Diagnostic In-
terview–Revised and Pre-Linguistic Autism Diagnostic
Observation Schedule/Autism Diagnostic Observation
Schedule, though at the time of this study the instru-
ments were distributed free of charge.
Funding/Support: This work was supported by grants
MH57167 and MH066469 from the National Institute of
Mental Health and HD 35482-01 from the National Insti-
tute of Child Health and Human Development (Dr Lord).
Disclaimer: This work was not written as part of Dr Thurm’s
official duties as a government employee. Views ex-
pressed in this article do not necessarily represent those
of the National Institutes of Health or the US government.
Previous Presentations: Parts of this work were pre-
sented at the Society for Research in Child Develop-
ment; April 23, 2003; Tampa, Fla; and April 17, 2001;
Minneapolis, Minn.
Acknowledgment: We thank D. Deborah Anderson, PhD,
Debra Combs, BA, E. Glenna Osborne, MA, Rebecca
Niehus, MA, Shanping Qiu, MA, and Lyn Carpenter, PhD,
for data collection and management assistance.
1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition, Text Revision. Washington, DC: American Psychiat-
ric Association; 2000.
2. World Health Organization. The ICD-10 Classification of Mental and Behavioral
Disorders: Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzer-
land: World Health Organization; 1992.
3. De Giacomo A, Fombonne E. Parental recognition of developmental abnormali-
ties in autism. Eur Child Adolesc Psychiatry. 1998;7:131-136.
4. Lord C. Follow-up of two-year-olds referred for possible autism. J Child Psychol
Psychiatry. 1995;36:1365-1382.
5. Moore V, Goodson S. How well does early diagnosis of autism stand the test of
time? follow-up study of children assessed for autism at age 2 and development
of an early diagnostic service. Autism. 2003;7:47-63.
6. Stone WL, Lee EB, Ashford L, Brissie J, Hepburn SL, Coonrod EE, Weiss BH.
Can autism be diagnosed accurately in children under 3 years? J Child Psychol
Psychiatry. 1999;40:219-226.
7. Gillberg C, Ehlers S, Schaumann H, Jakobsson G, Dahlgren SO, Lindblom R, Bagen-
holm A, Tjuus T, Blidner E. Autism under age 3 years: a clinical study of 28 cases
referred for autistic symptoms in infancy. J Child Psychol Psychiatry. 1990;
8. McGovern CW, Sigman M. Continuity and change from early childhood to ado-
lescence in autism. J Child Psychol Psychiatry. 2005;46:401-408.
9. Charman T, Taylor E, Drew A, Cockerill H, Brown JA, Baird G. Outcome at 7 years
of children diagnosed with autism at age 2: predictive validity of assessments
conducted at 2 and 3 years of age and pattern of symptom change over time.
J Child Psychol Psychiatry. 2005;46:500-513.
10. Lovaas OI. Behavioral treatment and normal educational and intellectual func-
tioning in young autistic children. J Consult Clin Psychol. 1987;55:3-9.
11. Strain PS. Generalization of autistic children’s social behavior change: effects of
developmentally integrated and segregated settings. Analysis Intervent Dev
Disabilities. 1983;3:23-34.
12. Sheinkopf SJ, Siegel B. Home-based behavioral treatment of young children with
autism. J Autism Dev Disord. 1998;28:15-23.
13. Smith T, Groen AD, Wynn JW. Randomized trial of intensive early intervention
for children with pervasive developmental disorder. Am J Ment Retard. 2000;
14. Wing L, Gould J. Severe impairments of social interaction and associated ab-
normalities in children: epidemiology and classification. J Autism Dev Disord.
15. Bailey A, Le Couteur A, Gottesman I, Bolton P, Simonoff E, Yuzda E, Rutter M.
Autism as a strongly genetic disorder: evidence from a British twin study. Psy-
chol Med. 1995;25:63-77.
16. Volkmar FR, Klin A, Siegal B, Szatmari P, Lord C, Campbell M, Freeman BJ, Cicchetti
DV, Rutter M, Kline W, Buitelaar J, Hattab Y, Fombonne E, Fuentes J, Werry J, Stone
W, Kerbeshian J, Hoshino Y, Bregman J, Loveland K, Szymanski L, Towbin K.
Field trial for autistic disorder in DSM-IV. Am J Psychiatry. 1994;151:1361-1367.
17. Buitelaar JK, Van der Gaag R, Klin A. Exploring the boundaries of pervasive de-
velopmental disorder not otherwise specified: analyses of data from the DSM-IV
autistic disorder field trial. J Autism Dev Disord. 1999;29:33-43.
18. Towbin KE. Pervasive developmental disorder not otherwise specified. In: Co-
hen DJ, Volkmar FR, eds. The Handbook of Autism and Other Pervasive Devel-
opmental Disorders. New York, NY: John Wiley & Sons; 1997:123-147.
19. Volkmar FR, Klin A, Schultz R, Bronen R, Marans WD, Sparrow S, Cohen DJ.
Asperger’s syndrome. J Am Acad Child Adolesc Psychiatry. 1996;35:118-123.
20. Gillberg C. Asperger syndrome and high-functioning autism. Br J Psychiatry. 1998;
21. Tanguay PE. Pervasive developmental disorders: a 10-year review. J Am Acad
Child Adolesc Psychiatry. 2000;39:1079-1095.
22. Cox A, Klein K, Charman T, Baird G, Baron-Cohen S, Swettenham J, Drew A, Wheel-
wright S. Autism spectrum disorders at 20 and 42 months of age: stability of
clinical and ADI-R diagnosis. J Child Psychol Psychiatry. 1999;40:719-732.
23. Sparrow S, Balla D, Cicchetti D. Vineland Adaptive Behavior Scales. Circle Pines,
Minn: American Guidance Service; 1984.
24. Mullen E. The Mullen Scales of Early Learning. Circle Pines, Minn: American Guid-
ance Service Inc; 1995.
25. Wechsler D. Wechsler Intelligence Scale for Children. 3rd ed. San Antonio, Tex:
Psychological Corp; 1991.
26. Elliott CD. Differential Ability Scales (DAS). San Antonio, Tex: Psychological Corp;
27. Lord C, Shulman C, DiLavore P. Regression and word loss in autistic spectrum
disorders. J Child Psychol Psychiatry. 2004;45:936-955.
28. Lord C, Rutter M, LeCouteur A. The Autism Diagnostic Interview–Revised: a re-
vised version of a diagnostic interview for caregivers of individuals with possible
pervasive developmental disorders. J Autism Dev Disord. 1994;24:659-685.
29. Risi S, Lord C, Corsello C, Chrysler C, Szatmari P, Cook EH, Leventhal BL, Pick-
les A. Combining information from multiple sources in the diagnosis of autism
spectrum disorders. J Am Acad Child Adolesc Psychiatry. In press.
30. Lord C, Risi S, Lambrecht L, Cook EH, Leventhal BL, DiLavore PC, Pickles A,
Rutter M. The Autism Diagnostic Observation Schedule–Generic: a standard mea-
sure of social and communication deficits associated with the spectrum of autism.
J Autism Dev Disord. 2000;30:205-223.
31. Lord C, Rutter ML, Goode S, Heemsbergen J, Jordan H, Mawhood L, Schopler
E. Autism Diagnostic Observation Schedule: a standardized observation of com-
municative and social behaviour. J Autism Dev Disord. 1989;19:185-212.
32. DiLavore PC, Lord C, Rutter M. The Pre-Linguistic Autism Diagnostic Observa-
tion Schedule. J Autism Dev Disord. 1995;25:355-379.
33. StataCorp. Stata Statistical Software, Release 8.0. College Station, Tex: Stata-
Corp; 2003.
34. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled
disagreement of partial credit. Psychol Bull. 1968;70:213-220.
35. Pickles A. Generalized estimating equations. In: Armitage P, Colton T, eds. The En-
cyclopedia of Biostatistics. Vol 2. New York, NY: John Wiley & Sons; 1998:1626-1637.
36. Hogan JW, Lancaster T. Instrumental variable and inverse probability weighting
for causal inference from longitudinal observational studies. Stat Methods Med
Res. 2004;13:17-48.
©2006 American Medical Association. All rights reserved.
at National Institute of Hlth, on March 27, 2008 www.archgenpsychiatry.comDownloaded from
... ASDs encompass a spectrum of neurological disorders related to an abnormal neurodevelopment process. These disorders affect individuals early in life, with clinical symptoms manifesting until the age of 5 years (Lord et al., 2006). The diagnosis, which is performed by clinical evaluation, follows international standards determined by the standards described in DSM-5 (Diagnostic and Statistical Manual of Mental Disorders) and takes into account deficiencies in social interaction, communication, limited interest, and repetitive behavior (Lord et al., 2006). ...
... These disorders affect individuals early in life, with clinical symptoms manifesting until the age of 5 years (Lord et al., 2006). The diagnosis, which is performed by clinical evaluation, follows international standards determined by the standards described in DSM-5 (Diagnostic and Statistical Manual of Mental Disorders) and takes into account deficiencies in social interaction, communication, limited interest, and repetitive behavior (Lord et al., 2006). ...
Full-text available
Neurological disorders include a wide spectrum of clinical conditions affecting the central and peripheral nervous systems. For these conditions, which affect hundreds of millions of people worldwide, generally limited or no treatments are available, and cell-based therapies have been intensively investigated in preclinical and clinical studies. Among the available cell types, mesenchymal stem/stromal cells (MSCs) have been widely studied but as yet no cell-based treatment exists for neurological disease. We review current knowledge of the therapeutic potential of MSC-based therapies for neurological diseases, as well as possible mechanisms of action that may be explored to hasten the development of new and effective treatments. We also discuss the challenges for culture conditions, quality control, and the development of potency tests, aiming to generate more efficient cell therapy products for neurological disorders.
... The ADI-R has been used in several studies to assess the evolution of patients [24,[30][31][32][33][34]. Lord and colleagues, for example, measured the ADI-R score evolution in 172 children followed-up with usual treatment between two assessments realized at two and 9 years of age [35]. They showed that the three mean ADI-R subscores remained stable, with only a few patients who were classified as having pervasive developmental disorder not otherwise specified (PDD-NOS) at 2 years old and became classified as autistic at the age of nine. ...
Full-text available
Background The Intensive, Interactive, and Individual (3i) intervention approach aims to decrease the severity of autism spectrum disorder (ASD) using intensive developmental play therapy (3i). We performed a retrospective study of 90 children who were enrolled for 2 years in the 3i approach to assess changes and predictors of changes in ASD severity at follow-up (FU). Methods The ASD severity of all patients (N = 119) who began 3i intervention between 2013 and 2018 was systematically measured using the childhood autism rating scale (CARS) and autism diagnosis interview-revised (ADI-R). Among them, 90 patients (mean age 5.6 ± 3.7 years) had a second assessment at the 2 year FU. CARS and ADI-R scores after 2 years of 3i intervention were compared to baseline scores using paired student’s t-tests. We used multiple linear regression models to assess the weight of baseline variables (e.g., age, oral language, sex, treatment intensity) on changes at the 2 year FU. Results Mean CARS and ADI-R subscores (interaction, communication, repetitive behaviour) decreased significantly by 20, 41, 27.5 and 25%, respectively (effect sizes: d > 0.8). Moreover, 55 and 46.7% of participants switched to a lower category of ASD severity based on the CARS scale and ADI-R interview, respectively. Multiple linear models showed that (i) a higher treatment intensity (more than 30 h per week) was significantly associated with a greater decrease (improvement) in the ADI-R interaction score; (ii) patients categorized as verbal subjects at baseline were associated with a better outcome, as ascertained by the CARS, ADI-R interaction and ADI-R communication scores; and (iii) older patients were significantly associated with a greater decrease in the ADI-R interaction score. However, we found no impact of sex, severity of ASD or comorbidities at baseline. Conclusion This study performed on 90 children suggests that 3i therapy may allow for a significant reduction in ASD severity with improvements in interaction, communication, and repetitive behaviours. A study using a control group is required to assess the efficacy of 3i play therapy compared to other interventions.
... Developmental quotient scores (DQ) were computed for each subdomain of the MSEL by dividing the individual developmental age by the chronological age and multiplying by 100 as described in 2006 by Lord et al. (46). The composite DQ was computed by calculating the average of all four subdomains' developmental ages, then dividing by the chronological age and multiplying by 100. ...
Full-text available
Evidence-based, early intervention significantly improves developmental outcome in young children with autism. Nonetheless, there is high interindividual heterogeneity in developmental trajectories during the therapy. It is established that starting intervention as early as possible results in better developmental outcomes. But except for younger age at start, there is no clear consensus about behavioral characteristics that could provide a reliable individual prediction of a child's developmental outcome after receiving an early intervention. In this study, we analyze developmental trajectories of preschoolers with autism who received 2 years of intervention using the Early Start Denver Model (ESDM) approach in Geneva, Switzerland in an individual setting (n = 55, aged 28.7 ± 5.1 months with a range of 15–42). Our aim was to identify early predictors of response to intervention. We applied a cluster analysis to distinguish between 3 groups based on their cognitive level at intake, and rates of cognitive change over the course of intervention. The first group of children only had a mild cognitive delay at intake and nearly no cognitive delay by the end of intervention (Higher Cognitive at baseline: HC). The children in the two other groups all presented with severe cognitive delay at baseline. However, they had two very different patterns of response to intervention. The majority significantly improved developmental scores over the course of intervention (Optimal Responders: OptR) whereas a minority of children showed only modest improvement (Minimal Responders: MinR). Further analyses showed that children who ended up having an optimal 2-year intervention outcome (OptR) were characterized by higher adaptive functioning at baseline combined with rapid developmental improvement during the first 6 months of intervention. Inversely, less significant progress by the sixth month of intervention was associated with a less optimal response to treatment (MinR).
... Variability in diagnostic age for ASD is evident, with children being diagnosed predominantly in the preschool years. Children may receive an early diagnosis, prior to the preschool years, as young as 2 years (Guthrie et al., 2013;Lord et al., 2006;Pierce et al., 2019;Wetherby et al., 2004), a later diagnosis at 6 or 7 years old (Mandell et al., 2005;Shattuck et al., 2009), or even in their adolescent years (Bargiela et al., 2016;Happé et al., 2016). Although the age of diagnosis varies across families of children with ASD, having a single consistent and valid social communication classification tool for children across the age range would enable families to utilize a common language with professionals (e.g., clinicians, therapists, teachers, etc.) who provide services for their children throughout the childhood years. ...
Full-text available
The Autism Classification System of Functioning: Social Communication (ACSF) describes social communication functioning levels. First developed for preschoolers with ASD, this study tests an expanded age range (2-to-18 years). The ACFS rates the child’s typical and best (i.e., capacity) performance. Qualitative methods tested parent and clinician perspectives of the ACSF age expansion using content analysis. The ACSF was used twice by parents and professionals for the same child/youth. Reliabilities were assessed using weighted kappa. Content validity supported the ACSF’s applicability, clarity, and usability. The ACSF adaptations did not change its original construct. Reliability were calculated from 90 parent and professional Time-1 and Time-2 ratings for children/youth (2.1–15.6 years). Results showed good-to-very good intra-rater agreement (typical) and good inter-rater agreement (capacity).
... Recent studies have explored the potential use of SAR in special education and specifically for individuals with cognitive disorders such as autism. Autism spectrum disorder (ASD) is a developmental disorder (DD) that characterizes children with a set of eccentric patterns of behavior that appears in infant or toddler years [2], while the deficits fall into three categories: social interaction, communication, and stereotyped behavior [3]. Furthermore, due to the absence of permanent cure, specialized intervention plans are taking place to reduce the autistic behavior of children, while producing significant results [4]. ...
Full-text available
Socially assistive robots are widely deployed in interventions with children on the autism spectrum, exploiting the benefits of this technology in social behavior intervention plans, while reducing their autistic behavior. Furthermore, innovations in modern technologies such as machine learning enhance these robots with great capabilities. Since the results of this implementation are promising, their total cost makes them unaffordable for some organizations while the needs are growing progressively. In this paper, a low-cost robot for autism interventions is proposed, benefiting from the advantages of machine learning and low-cost hardware. The mechanical design of the robot and the development of machine learning models are presented. The robot was evaluated by a small group of educators for children with ASD. The results of various model implementations, together with the design evaluation of the robot, are encouraging and indicate that this technology would be advantageous for deployment in child–robot interaction scenarios.
... Clinical best estimate of autism. A clinical best estimate (CBE) procedure was developed from the CBE procedures of Lord and colleagues (Lord et al., 2006;Lord, Petkova, et al., 2012) to determine autism diagnosis based on a review of the Autism Diagnostic Observation Schedule -2nd edition (ADOS-2; Lord, Rutter, et al., 2012), MSEL, and the Vineland Adaptive Behavior Scales (Sparrow et al., 2005) combined with clinical expertise in autism (see Hogan et al., 2017). The CBE was conducted by a team, including a licensed psychologist, who were all research reliable on the ADOS-2 (Hogan et al., 2017;Roberts et al., 2020). ...
Background and Aims Individuals with fragile X syndrome (FXS) characteristically struggle with language and communication throughout the life course, but there is limited research on the development of communication before 24 months. The purpose of this study is to describe the early communication of infants and toddlers with FXS using the Communication and Symbolic Behavior Scales-Caregiver Questionnaire (CSBS-CQ), a standardized communication screening measure, as compared to the reported normative data of the CSBS-CQ and identify the percentage of infants and toddlers who scored within the range of concern. Documenting how children with FXS perform on screening measures can provide a quick snapshot of skills to help clinicians determine the need for services. Methods Participants were 22 infants and toddlers with FXS between 6 and 29 months. Performance on the CSBS-CQ was compared to the measure's normative data. The CSBS-CQ was completed by mothers, and children were administered the Mullen Scales of Early Learning. Because co-occurring autism is common in FXS, the presence of autism was determined using a clinical best estimate procedure. Results Overall and within the domains and subdomains of the CSBS-CQ, infants and toddlers with FXS had significantly lower scores than the normative data. Further, 68.2% of our sample was in the range of concern for their overall communication score. The presence of autism led to consistently lower scores, and more infants and toddlers with FXS + autism scored within the range of concern. Conclusions Our findings suggest that delays in early communication are evident in comparison to typically developing norms before 24 months. These findings also emphasize that infants and toddlers with FXS would likely benefit from early language intervention given that 68.2% of our sample was in the range of concern for their overall communication score. Implications Early identification and developmental monitoring of children with FXS will help to determine concerns in communication and other domains of development. While early communication broadly may not be an early indicator of autism in FXS, some specific skills, such as eye gaze, may serve as such an indicator. Screening measures, like the CSBS-CQ, may help monitor both early communication impairments and autism symptoms. Infants and toddlers with FXS, regardless of autism status, will benefit from early language interventions.
... Early intervention is critical to supporting infants with ASD (Jones et al., 2014) and the importance of early diagnosis is recognised in a recent UK Government strategy (UK Government, 2021). However, high stability and reliability of autism diagnosis has been found only to occur from the age of 2-3 years when behavioural signs begin to present (Cox et al., 1999;Lord, 1995;Lord et al., 2006;Moore & Goodson, 2003;Stone et al., 1999). ...
This thesis describes the development and application of age-appropriate structural priors to improve the localisation accuracy of diffuse optical tomography (DOT) approaches in infants aged from birth to two years of age. Knowledge of the target cranial anatomy, known as a structural prior, is required to produce three-dimensional images localising concentration changes to the cortex. A structural prior would ideally be subject-specific, i.e. derived from structural magnetic resonance imaging (MRI) data from each specific subject. Requiring a structural scan from every infant participant, however, is not feasible and undermines many of the benefits of DOT. A review was conducted to catalogue available infant structural MRI data, and selected data was then used to produce structural priors for infants aged 1- to 24-months. Conventional analyses using functional near-infrared spectroscopy (fNIRS) implicitly assume that head size and array position are constant across infants. Using DOT, the validity of assuming these parameters constant in a longitudinal infant cohort was investigated. The results show that this assumption is reasonable at the group-level in infants aged 5- to 12-months but becomes less valid for smaller group sizes. A DOT approach was determined to illicit more subtle effects of activation, particularly for smaller group sizes and expected responses. Using state-of-the-art MRI data from the Developing Human Connectome Project, a database of structural priors of the neonatal head was produced for infants aged pre-term to term-equivalent age. A leave-one-out approach was used to determine how best to find a match between a given infant and a model from the database, and how best to spatially register the model to minimise the anatomical and localisation errors relative to subject-specific anatomy. Model selection based on the 10/20 scalp positions was determined to be the best method (of those based on external features of the head) to minimise these errors.
Background A formal autism diagnosis can be an inefficient and lengthy process. Families may wait several months or longer before receiving a diagnosis for their child despite evidence that earlier intervention leads to better treatment outcomes. Digital technologies that detect the presence of behaviors related to autism can scale access to pediatric diagnoses. A strong indicator of the presence of autism is self-stimulatory behaviors such as hand flapping. Objective This study aims to demonstrate the feasibility of deep learning technologies for the detection of hand flapping from unstructured home videos as a first step toward validation of whether statistical models coupled with digital technologies can be leveraged to aid in the automatic behavioral analysis of autism. To support the widespread sharing of such home videos, we explored privacy-preserving modifications to the input space via conversion of each video to hand landmark coordinates and measured the performance of corresponding time series classifiers. Methods We used the Self-Stimulatory Behavior Dataset (SSBD) that contains 75 videos of hand flapping, head banging, and spinning exhibited by children. From this data set, we extracted 100 hand flapping videos and 100 control videos, each between 2 to 5 seconds in duration. We evaluated five separate feature representations: four privacy-preserved subsets of hand landmarks detected by MediaPipe and one feature representation obtained from the output of the penultimate layer of a MobileNetV2 model fine-tuned on the SSBD. We fed these feature vectors into a long short-term memory network that predicted the presence of hand flapping in each video clip. Results The highest-performing model used MobileNetV2 to extract features and achieved a test F1 score of 84 (SD 3.7; precision 89.6, SD 4.3 and recall 80.4, SD 6) using 5-fold cross-validation for 100 random seeds on the SSBD data (500 total distinct folds). Of the models we trained on privacy-preserved data, the model trained with all hand landmarks reached an F1 score of 66.6 (SD 3.35). Another such model trained with a select 6 landmarks reached an F1 score of 68.3 (SD 3.6). A privacy-preserved model trained using a single landmark at the base of the hands and a model trained with the average of the locations of all the hand landmarks reached an F1 score of 64.9 (SD 6.5) and 64.2 (SD 6.8), respectively. Conclusions We created five lightweight neural networks that can detect hand flapping from unstructured videos. Training a long short-term memory network with convolutional feature vectors outperformed training with feature vectors of hand coordinates and used almost 900,000 fewer model parameters. This study provides the first step toward developing precise deep learning methods for activity detection of autism-related behaviors.
Full-text available
Objectives Autism spectrum disorder (ASD) is a neurodevelopmental condition marked by stereotyped behavior and poor social interaction. Although the etiology of this illness is unknown, research clearly shows that it has a genetic foundation due to complicated inheritance. It affects about 52 million individuals worldwide. Several risk factors for autism converge into possible pathways for other neurodevelopmental diseases, with onsets occurring at various stages of development. Methods In the study’s literature review, the genes included were identified in articles published over the previous 30 years in databases such as the web of sciences, PubMed, Google Scholar, Embase, and other databases. Candidate genes associated with ASD are CHD8, SHANK3, SLC6A4, RELN, DISC1, and ITGB3. Results Several prenatal risk factors cause neurological vulnerability, which increases the probability of autism and other neurodevelopmental problems. Genomic research has allowed tremendous progress in discovering ASD risk genes during the last decade. Recent technological advancements have demonstrated that certain genetic mutations and modifications may serve as useful biological markers, risk indicators, and therapeutic targets for illnesses. Conclusions In large cohorts, high-throughput next-generation sequencing uncovers a varied and complicated genetic landscape of new risk genes. More studies are needed to understand better the environmental variables that play a crucial role in disease development. Currently, there is less clinical data to support the function of ASD. However, the prevailing research facts for many researched ASD new candidate genes support their links and identify ASD etiologic processes for establishing an early diagnostic marker.
Autism spectrum disorder (ASD) is a prevalent and poorly understood neurodevelopmental disorder. There are currently no laboratory-based diagnostic tests to detect ASD, nor are there any disease-modifying medications that effectively treat ASD’s core behavioral symptoms. Scientific progress has been impeded, in part, by overreliance on model organisms that fundamentally lack the sophisticated social and cognitive abilities essential for modeling ASD. We therefore saw significant value in studying naturally low-social rhesus monkeys to model human social impairment, taking advantage of a large outdoor-housed colony for behavioral screening and biomarker identification. Careful development and validation of our animal model, combined with a strong commitment to evaluating the translational utility of our preclinical findings directly in patients with ASD, yielded a robust neurochemical marker (cerebrospinal fluid vasopressin concentration) of trans-primate social impairment and a first-in-class medication (intranasal vasopressin) shown in a small phase 2a pilot trial to improve social abilities in children with ASD. This translational research approach stands to advance our understanding of ASD in a manner not readily achievable with existing animal models, and can be adapted to investigate a variety of other human brain disorders which currently lack valid preclinical options, thereby streamlining translation and amplifying clinical impact more broadly.
Full-text available
The Autism Diagnostic Observation Schedule—Generic (ADOS-G) is a semistructured, standardized assessment of social interaction, communication, play, and imaginative use of materials for individuals suspected of having autism spectrum disorders. The observational schedule consists of four 30-minute modules, each designed to be administered to different individuals according to their level of expressive language. Psychometric data are presented for 223 children and adults with Autistic Disorder (autism), Pervasive Developmental Disorder Not Otherwise Specified (PDDNOS) or nonspectrum diagnoses. Within each module, diagnostic groups were equivalent on expressive language level. Results indicate substantial interrater and test—retest reliability for individual items, excellent interrater reliability within domains and excellent internal consistency. Comparisons of means indicated consistent differentiation of autism and PDDNOS from nonspectrum individuals, with some, but less consistent, differentiation of autism from PDDNOS. A priori operationalization of DSM-IV/ICD-10 criteria, factor analyses, and ROC curves were used to generate diagnostic algorithms with thresholds set for autism and broader autism spectrum/PDD. Algorithm sensitivities and specificities for autism and PDDNOS relative to nonspectrum disorders were excellent, with moderate differentiation of autism from PDDNOS.
Nosology Conceptual Background Biological Studies Of PDD-NOS Differential Diagnosis Epidemiology Etiology Natural History Treatment Summary
Twenty children who presented with severe interactional and communication difficulties at age 2 underwent a comprehensive assessment for autism, and were reassessed at age 4-5. In common with other recent studies, diagnosis of autistic spectrum disorders at age 2 was found to be reliable and stable. The communication and social skills of the children showed little change overall by the second assessment. However, children whose scores deteriorated in the social domain tended to have presented initially with more significant behaviour problems. Few repetitive behaviours were observed at age 2, whereas these were more apparent by age 4-5. The finding that early diagnosis of autism is reliable and stable has led to the development of an early diagnostic service in Southampton, which is described. The importance of early diagnosis is that it opens the door to early intervention programmes, which in turn prevent many problems from occurring in later life.
The association between, and stability of, clinical diagnosis and diagnosis derived from the Autism Diagnostic Interview-Revised (ADI-R; Lord, Rutter, & Le Couteur, 1994) was examined in a sample of prospectively identified children with childhood autism and other pervasive developmental disorders assessed at the age of 20 months and 42 months. Clinical diagnosis of autism was stable, with all children diagnosed with childhood autism at age 20 months receiving a diagnosis of childhood autism or a related pervasive developmental disorder (PDD) at age 42 months. Clinical diagnosis of childhood autism was also reasonably sensitive, with all children who went on to receive a clinical diagnosis of childhood autism at 42 months being identified as having autism or PDD at 20 months. However, clinical diagnosis for PDD and Asperger's syndrome lacked sensitivity at 20 months, with several children who subsequently received these diagnoses at 42 months receiving diagnoses of language disorder or general developmental delay, as well as in two cases being considered clinically normal, at the earlier timepoint. The ADI-R was found to have good specificity but poor sensitivity at detecting childhood autism at 20 months; however, the stability of diagnosis from 20 to 42 months was good. In addition, the ADI-R at age 20 months was not sensitive to the detection of related PDDs or Asperger's syndrome. The continuity and discontinuity between behavioural abnormalities identified at both timepoints in the three domains of impairment in autism was examined, both in children who met final clinical criteria for an autistic spectrum disorder, and for children with language disorder who did not, as well as for a small sample of typically developing children.
Longitudinal data occurs frequently in medical studies, particularly in clinical trials. Often, the response variable is nonnormal, for example, it may be a binary variable, that is, “improved” or “not improved”. The approach to the analysis of such data discussed in this entry is the use of generalized estimating equations, which, while making weaker distributional assumptions than those required for a fully likelihood-based model, maintains the properties of consistency and asymptotic normality of parameter estimates. Keywords: longitudinal data; working correlation matrix; sandwich estimator; quasi-likelihood
This study was specifically designed to test the notion that a developmentally-integrated setting would yield superior generalized behavior change than would a developmentally-segregated setting. Four autistic boys served as target subjects. Each day, three 20-min play sessions were conducted. One setting was devoted to peer-mediated training, one to integrated generalization assessment, and one to segregated generalization assessment. The order of the three sessions was counterbalanced across the days of the study. The study employed a multiple baseline design across subjects to demonstrate experimental control over the subjects' positive social interaction, and a combined multiple baseline and simultaneous treatment design to evaluate the impact of developmentally-segregated and developmentally-integrated settings on generalized behavior change. During the Baseline condition each boy engaged in consistently low levels of positive interaction during all sessions. During the Peer Social Initiations I. condition each boy was exposed sequentially to a peer-mediated treatment package. Each day of this condition an integrated and a segregated generalization session was conducted. Only when the boys were exposed to the intervention did their level of positive interaction increase during training sessions. In the final condition, Peer Social Initiations II., treatment continued without alteration. However, now both generalization sessions were integrated. For each boy, clearly superior generalization effects were obtained during integrated sessions. The results of this study have clear social policy implications vis a vis Public Law 94-142 and its controversial stipulation for the placement of handicapped children in the least restrictive environment.